Xuedong Huang

Intent, Knowledge, and Multisensory via Machine Learning

Online services equipped with devices of multisensory capabilities (touch, voice, vision, accelerometer, GPS…) are changing the way people live, work, and play. The underlying computational framework that makes use of massive amount of user activity data to conduct efficient intent-knowledge matching services is central to enable natural user interaction and efficient user task completion. The state of the art web search engine such as Bing and Google is only a baby step towards building such a powerful service to help people accomplish tasks in a much more natural and efficient way. The advancement of cloud computing and smart devices made it possible for us to employ computationally efficient machine learning technologies to perform such intent-knowledge matching services. This talk is illustrated with historical lessons learnt and is highlighted with an exciting vision for us to work together to fulfill the collective dream of generations.

Biography: http://www.microsoft.com/presspass/exec/de/Huang/default.mspx